Page 1 of 1
Confidence interval
Posted: Tue Oct 20, 2009 11:37 am
by AKM
I am a newbie to eViews. I've been running regressions but I am unable to find confidence intervals to the estimators. e.g. y=mx+c. After estimating m and c, I am not sure how to find a c.i for m = in excel you get to select the interval while regressing.
Any info is appreciated.
Regards,AKM
Re: Confidence interval
Posted: Thu Oct 22, 2009 9:47 am
by trubador
Confidence interval is an additional output and its calculation is fairly straightforward. In EViews, a typical output already gives you standard errors of the estimators. All you have to do is generate a series using this information. Suppose that the estimated coefficients c(1) and c(2) refer to c and m respectively. Then you can generate %95 upper and lower confidence bounds as follows:
Code: Select all
series upper_c1 = @coefs(1)+@qnorm(.025)*@stderrs(1)
series upper_c2 = @coefs(2)+@qnorm(.025)*@stderrs(2)
series lower_c1 = @coefs(1)-@qnorm(.025)*@stderrs(1)
series lower_c2 = @coefs(2)-@qnorm(.025)*@stderrs(2)
Re: Confidence interval
Posted: Thu Oct 22, 2009 9:57 am
by EViews Gareth
EViews 7 has the ability to show confidence intervals without any extra calculations.
Re: Confidence interval
Posted: Thu Apr 28, 2016 8:05 am
by wardy94
Hi there,
I was wondering how to compute confidence intervals for impulse response functions under VARs?
Thanks,
James